Compared to the conventional trench MOSFETs, super-junction trench MOSFETs are more attractive due to higher breakdown voltage and lower specific Rds (drain-source resistance). As is known to all, a super-junction trench MOSFET is implemented by a p type column structure and an n type column structure arranged in parallel and connecting to each other onto a heavily doped substrate, however, the manufacturing yield is not stable because the super-junction trench MOSFET is very sensitive to the fabrication processes and conditions such as: the p type column structure and the n type column structure dopant re-diffusion issue induced by subsequent thermal processes; trapped charges within the column structure, etc. . . . . All that will cause a hazardous condition of charges imbalance to the super-junction trench MOSFET. More specifically, these undesired influences become more pronounced with a narrower column structure width for a lower bias voltage ranging under 200V.
Prior art (paper “Industrialization of Resurf stepped oxide technology for Power Transistor”, by M. A. Gajda, etc., and paper “Tunable Oxide-Bypassed Trench Gate MOSFET Breaking the Ideal Super-junction MOSFET Performance Line at Equal Column Width”, by Xin Yant, etc.) disclosed device structures in order to resolve the limitation caused by the conventional super-junction trench MOSFET discussed above, as shown in FIG. 1A and FIG. 1B. Except some different terminologies (the device structure in FIG. 1A named with RSO: Resurf Stepped Oxide and the device structure in FIG. 1B named with TOB: Tuable Oxide-Bypassed). Both the device structures in FIG. 1A and FIG. 1B are basically the same which can achieve a lower specific Rds and a higher breakdown voltage than a conventional super-junction trench MOSFET because each the epitaxial layer formed in FIG. 1A and FIG. 1B has a higher doping concentration than the conventional super-junction trench MOSFET.
Refer to FIG. 1A and FIG. 1B again, both the device structures have a deep trench with a thick oxide layer along trench sidewalls and bottoms into a drift region. Only difference is that, the device structure in FIG. 1A has a single epitaxial layer (N Epi, as illustrated in FIG. 1A) while the device structure in FIG. 1B has double epitaxial layers (Epi1 and Epi2, as illustrated in FIG. 1B, the Epi1 supported on a heavily doped substrate has a lower doping concentration than the Epi2 near a channel region.) Due to the p type column structure and the n type column structure inter-diffusion, both the device structures in FIG. 1A and FIG. 1B do not have charges imbalance issue, resolving the technical limitation caused by the conventional super-junction trench MOSFET, however, the benefit of both the device structures in FIG. 1A and FIG. 1B over the conventional super-junction trench MOSFET only pronounces at the bias voltage ranging under 200V, which means that, the conventional super-junction trench MOSFET has a lower Rds when the bias voltage is beyond 200V.
U.S. Pat. No. 7,601,597 disclosed a method to avoid the aforementioned p type column structure and the n type structure dopant re-diffusion issue, for example in an N-channel trench MOSFET as shown in FIG. 1C, by setting up the p type column formation process at a last step after all diffusion processes such as: sacrificial oxidation after trench etch, gate oxidation, P body region formation and n+ source region formation, etc. . . . have been finished.
However, the disclosed method of this prior art is not effective because that, firstly, the p type column structure is formed by growing an additional p type epitaxial layer in deep trenches etched in an n type epitaxial layer; secondly, an additional CMP (Chemical Mechanical Polishing) is required for surface planarization after the p type epitaxial layer is grown; thirdly, double trench etches are necessary (one for shallow trenches for trenched gates formation and another for the deep trenches for the p type column structure formation), all the increased cost is not conductive to mass production. Moreover, other factors such as: the charges imbalance caused by the trapped charges within the column structure is still not resolved.
Therefore, there is still a need in the art of the semiconductor power device, particularly for super-junction trench MOSFET design and fabrication, to provide a novel cell structure, device configuration that would resolve these difficulties and design limitations.